考虑情感强度的加权社会网络偏好信息识别研究

Weighted social network preference information recognition considering emotional intensity

为了准确的识别网络文本中用户的情感偏好信息,提出了一种考虑情感强度的加权社会网络偏好信息识别算法。首先提取出文本特征项的独立信息,使用词典与互信息相结合的分词方法对文本做分词处理。分析实际文本中副词可以表达出的情感强度,将不同情感强度的副词赋予不同权重值,通过将句子本身定义的权重值与句中副词权值相乘来获得文本总情感强度。对提取出的特征项做向量转化处理,通过GMM算法进行情感偏好状态测定,完成识别全过程。仿真分析实验表明,本文算法可行性较强,识别效果较优。

In order to accurately identify the user’s emotional preference information in the network text,a weighted social network preference information recognition considering emotional intensity is proposed.First,we extract the independent information of the text features,and use the word segmentation method which combines dictionary and mutual information to segment the text.Analyze the emotional intensity that can be expressed by adverbs in the actual text,assign adverbs with different emotional intensity with different weight values,and obtain the total emotional intensity of the text by multiplying the weight value defined by the sentence itself and the weight value of adverbs in the sentence.Finally,the extracted feature items are transformed into vectors,and GMM algorithm is used to measure the emotion preference state to complete the whole process of recognition.The simulation results show that the algorithm is feasible and the recognition effect is better.